关键词: 数据挖掘, 分布式数据库, 安全, 隐私

Abstract: Abstract Fast mining of associate rules and global frequent itemsets is an important problem in a distributed database environment. But it is crucial to preserve user’s privacy without disclosing individual transactions. This paper addresses an algorithm PPDMA (Privacy Preserving Distributed Mining) for association rules over horizontally partitioned data. The methods apply cryptographic techniques to minimize the information shared. Theory analysis shows the feasibility and effectiveness of the algorithm.

Key words: data mining, distributed database, security, privacy